The 1st-place Solution for ECCV 2022 Multiple People Tracking in Group Dance Challenge

TL;DR

This paper presents a top-performing solution for multiple people tracking in group dance videos, leveraging transformer-based detection and tracking techniques, joint training, and YOLOX proposals, achieving state-of-the-art accuracy.

Contribution

The authors introduce a novel transformer-based tracking method with query denoising and pseudo video training, significantly improving tracking performance in dance scenarios.

Findings

Achieved 73.4% HOTA on DanceTrack test set

Surpassed second-place by 6.8% HOTA

Demonstrated effectiveness of YOLOX proposals in tracking

Abstract

We present our 1st place solution to the Group Dance Multiple People Tracking Challenge. Based on MOTR: End-to-End Multiple-Object Tracking with Transformer, we explore: 1) detect queries as anchors, 2) tracking as query denoising, 3) joint training on pseudo video clips generated from CrowdHuman dataset, and 4) using the YOLOX detection proposals for the anchor initialization of detect queries. Our method achieves 73.4% HOTA on the DanceTrack test set, surpassing the second-place solution by +6.8% HOTA.